ferrox-plan-review-convergence

Repeat plan revision and external review cycles until high-priority concerns converge.

21|9|Updated Jul 21, 2026
One-click install
npx skills add https://github.com/FerroxLabs/ferrox-factory --skill ferrox-plan-review-convergence
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ferrox-plan-review-convergence
Source: https://github.com/FerroxLabs/ferrox-factory/tree/main/skills/ferrox-plan-review-convergence
Command: npx skills add https://github.com/FerroxLabs/ferrox-factory --skill ferrox-plan-review-convergence

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents plans from stagnating after review by forcing repeated revision cycles until unresolved high-priority concerns are actually addressed. It is designed for workflows where a plan must survive external critique, not merely accumulate comments.

Core Features & Use Cases

  • Cross-AI review orchestration: Runs plan-phase, then sends the result through external reviewers such as Codex, Gemini, Claude, OpenCode, or local model servers.
  • Replan-and-recheck loop: Feeds review feedback back into the planning phase and repeats until the plan converges or the cycle limit is reached.
  • Workflow gates and escalation: Preserves pre-flight validation, stall detection, and escalation handling so unresolved issues do not slip through.
  • Use case: A complex feature plan keeps coming back with unresolved architectural risks, and this Skill is used to iterate until the plan is genuinely ready to execute.

Quick Start

Run the ferrox-plan-review-convergence skill for phase 2 using your preferred reviewer flags and a max cycle limit.

Frequently Asked Questions about ferrox-plan-review-convergence

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate a plan revision loop with cross-AI review until all concerns are resolved?

You can automate plan revision loops by running the plan through external reviewers and feeding feedback back into replanning. This cycle repeats until unresolved high-priority concerns are eliminated or a cycle limit is reached.

What is the best way to use multiple AI reviewers like Codex, Gemini, or Claude for software plan validation?

Using multiple AI reviewers for software plan validation involves orchestrating plan-phase output through external models. Reviewer selection directs the plan to local or CLI-based providers like Codex, Gemini, or Claude for critique.

How do I prevent planning workflows from stalling after an AI review identifies architectural risks?

To prevent planning workflows from stalling, you can apply stall detection and escalation gates. These mechanisms force replanning cycles until architectural risks and high-priority concerns are genuinely addressed.

How does convergence logic handle unresolved high-priority concerns in a phased software plan?

Convergence logic handles unresolved concerns by applying escalation gates that block execution. It forces repeated replanning and review cycles until the plan converges or a defined maximum cycle limit is hit.

Do I need pre-flight validation before running an automated cross-AI replanning and review cycle?

Yes, pre-flight validation is preserved within the workflow before initiating the revision loop. It ensures the phased software plan meets baseline requirements before sending it to external AI reviewers.